unknaszero option

[UstasAlex]

When SNMP query doesn't return anything, Cacti draws nothing and thats fine, but when SNMP starts working again Cacti fills the empty space on the graph as if there was no problem. Is there any way to fix it?
in MRTG for instance there is a "unknaszero option". Does Cacti have anything liek that?

[SecuringStuff]

Hi UstasAlex,

Please forgive me if I did not understand your question correctly.

Cacti shows the unknown data in gray color and if your cacti does not then you can add it to the graph template.

it seems to me that gray area or 0 line/area are the same.

Regards,

[UstasAlex]

Hi UstasAlex,

Please forgive me if I did not understand your question correctly.

Cacti shows the unknown data in gray color and if your cacti does not then you can add it to the graph template.

it seems to me that gray area or 0 line/area are the same.

Regards,

I will try to describe the problem Im facing in more details.

I have Cacti and MRTG installed on freebsd box. They both monitor the number of VPDN sessions on cisco router. They query the router every minute.

Yesterday I accidently wrote an ACL that denied all SNMP traffic on the router. That ACL was on for 15 minutes. Then I corrected the ACL.

But when I looked at the Cacti and MRTG grapgh I saw that there was no indication of the fact that the router was unreachable for 15 minutes.
both Cacti and MRTG showed graphs as if there was no problem.

After reading MRTG FAQ I found out that mrtg tends to draw a number from last susccessful SNMP response. so for instance when I turned SNMP down on the router I had 1544 sessions. and MRTG kept drawing this number for 15 minutes until I corrected ACL. Cacti probably did the same thing.
So MRTG FAQ says that by using "unknaszero option" the number can be switched to zero if no SNMP response from the queried device comes.
I used that option and denied SNMP once again and saw that it did work for MRTG. so I could see on my mrtg graph for how long the router was unreachable by SNMP. How can I do the same thing for Cacti?

[SecuringStuff]

I got your point.

I am not an expert but I only have suggestions so forgive me if my suggestion is not so helpful.

I think that you might want to change the "Failure Count" and "Recovery Count" under Poller settings to "1"

Regards,

[BSOD2600]

When SNMP query doesn't return anything, Cacti draws nothing and thats fine, but when SNMP starts working again Cacti fills the empty space on the graph as if there was no problem. Is there any way to fix it?
in MRTG for instance there is a "unknaszero option". Does Cacti have anything liek that?

1) Post a picture of the graph with this occuring.

2) It's rrdtool which does the graphing, NOT cacti. Read up on how rrdtool functions and it's options.

[gandalf]

Does Cacti have anything liek that?

No and yes.
It does NOT have it as a default CDEF, but you may simply add this CDEF and apply it to all grph items you want.
Reinhard

[UstasAlex]

Does Cacti have anything liek that?

No and yes.
It does NOT have it as a default CDEF, but you may simply add this CDEF and apply it to all grph items you want.
Reinhard

could you please explain what CDEF is and how I apply it?

[SecuringStuff]

could you please explain what CDEF is and how I apply it?

Hi,

Please read http://www.cacti.net/downloads/docs/html/cdefs.html

Regards,